langchain-pdf-qa

Use langchain to create a model that returns answers based on online PDFs that have been read.

How to use

1. Clone the repository and install dependencies

git clone git@github.com:m-star18/langchain-pdf-qa.git
pip install -r requirements.txt

2. Specify the PDF link and OPEN_API_KEY to create the embedding model

# Example
python src/make_emb.py --pdf_path "https://arxiv.org/pdf/2005.14165.pdf" --OPEN_API_KEY ""

The following options can also be specified as arguments

• chunk_size: Please specify the chunk_size for CharacterTextSplitter within a number less than or equal to 4096.
• chunk_overlap: Please specify the chunk_overlap for CharacterTextSplitter within a number less than or equal to 4096.
• split_mode: Please specify the split mode. (character, recursive_character, nltk, tiktoken)
• faiss_save_path: Please specify the name of the created Faiss object.

3. Enter a question and generate an answer from the extracted text

# Example
python src/pdf_qa.py --query "On which datasets does GPT-3 struggle?"